Technological Shocks and Stock Market Volatility Over a Century: A GARCH-MIDAS Approach
Research report
TLDR
Paper proposes a modified GARCH model that is able to account for technological shocks (TS) when forecasting volatility. Analysis on the long and short term impacts on volatility of TS show that it is possible to account for the effects of such shocks via a modified GARCH mixed data sampling (MIDAS) approach.
Key learnings
Accounting for tech shocks is important in traditional markets and are well-modeled by MIDAS models. Tech shocks are large innovations or new IP that change industry.
Average time for tech shock-induced volatility to be realized is 3-5 years. The lag between tech shocks and volatility is interesting to note, as crypto innovations take much less time to be impactful.
Methods and outputs
Paper first establishes a positive relationship between the occurrence of tech shocks (in the form of patents with a 5-year rolling window) and subsequently increased volatility. This is quantified and shown in Table 1 - page 28.
Primary output is a modified GARCH-MIDAS model aimed at forecasting volatility while accounting for tech shocks. Formulae for this model are provided on pages 9-12. It is unclear which packages/libraries the authors use, but the equations are relatively straight forward.
GARCH models are well established – refer here for more information. MIDAS models are well-suited for modeling high frequency dependent variables which exhibit conditional heteroskedasticity and a low-frequency predictor variable. As such, they are well-suited for modeling TS.
Testing is conducted on US-specific data (patents primarily) and OECD nations at large, both in and out of sample. The results show that the model yield better Sharpe ratios when applied to portfolio optimization and reduces volatility.
Concrete
Crypto is not an industry that lends itself to a slow rate of innovation, and that it is very difficult to distinguish signal from noise in the realm of crypto research. As such, this paper serves as an interesting read but yields little immediate, practical utility for Concrete’s purposes.
Links and further reading
Last updated